Good morning. My employer uses appd on all our java app servers, I work in the operations dept, and recently I discovered a rather disturbing trend. Way too many of our servers are getting tickets for disk utilization being above 84%, and it turns out the root cause is that the appd_agent service is holding open dozens or hundreds of huge deleted application log files in /var/log. Bouncing the appd_agent service has now become my first action whenever presented with high disk utilization. It seems to be more common on hosts running Apache Camel, as the busier hosts can roll over logfiles in a matter of seconds.
I don't know what appd_agent is doing with those logfiles. I cannot imagine a reason that a performance monitoring service would need to ever open an application logfile. Looking for errors, I'm sure. It may merely be a local configuration issue, causing those files to be scanned, but I'm not in the loop as to how appd is configured.
